(GST-3) and ten Delta-class (GST-1)

Biochem. J. (2003) 370, 661–669 (Printed in Great Britain) 661Cloning, expression and biochemical characterization of one Epsilon-class(GST-3) and ten Delta-class (GST-1) glutathione S-transferases fromDrosophila melanogaster, and identification of additional nine membersof the Epsilon classRafa SAWICKI*, Sharda P. SINGH*, Ashis K. MONDAL*, Helen BENES † and Piotr ZIMNIAK*‡§ 1*Department of Medicine, University of Arkansas for Medical Sciences, Little Rock, AR 72205, U.S.A., †Department of Anatomy and Neurobiology, University ofArkansas for Medical Sciences, Little Rock, AR 72205, U.S.A., ‡Department of Biochemistry and Molecular Biology, University of Arkansas for Medical Sciences, LittleRock, AR 72205, U.S.A., and §Central Arkansas Veterans Healthcare System, Medical Research (151/LR), 4300 West 7th Street, Little Rock, AR 72205, U.S.A.From the fruitfly, Drosophila melanogaster, ten members of thecluster of Delta-class glutathione S-transferases (GSTs; formerlydenoted as Class I GSTs) and one member of the Epsilon-classcluster (formerly GST-3) have been cloned, expressed inEscherichia coli, and their catalytic properties have been determined.In addition, nine more members of the Epsilon clusterhave been identified through bioinformatic analysis but notfurther characterized. Of the 11 expressed enzymes, seven acceptedthe lipid peroxidation product 4-hydroxynonenal assubstrate, and nine were active in glutathione conjugation of 1-chloro-2,4-dinitrobenzene. Since the enzymically active proteinsincluded the gene products of DmGSTD3 and DmGSTD7 whichwere previously deemed to be pseudogenes, we investigated themfurther and determined that both genes are transcribed inDrosophila. Thus our present results indicate that DmGSTD3and DmGSTD7 are probably functional genes. The existence andmultiplicity of insect GSTs capable of conjugating 4-hydroxynonenal,in some cases with catalytic efficiencies approachingthose of mammalian GSTs highly specialized for this function,indicates that metabolism of products of lipid peroxidation is ahighly conserved biochemical pathway with probable detoxificationas well as regulatory functions.Key words: detoxification, Diptera, electrophile, glutathioneconjugation, lipid peroxidation.INTRODUCTIONPartially reduced forms of oxygen, including the superoxideradical anion, hydrogen peroxide and the hydroxyl radical, arecollectively known as reactive oxygen species (ROS). The formationof ROS is an inevitable consequence of aerobic metabolism.The majority of ROS is generated by mitochondria, althoughother enzyme systems also contribute; transition metal ionscatalyse interconversions of individual ROS, especially the formationof the highly reactive hydroxyl radical. The latter can react,among other targets, with polyunsaturated fatty acids and thusinitiate the lipid peroxidation chain reaction [1], which producesmultiple molecules of lipid hydroperoxides per initiating event.By either spontaneous or enzyme-catalysed reactions [2], lipidhydroperoxides give rise to α,β-unsaturated aldehydes such as 4-hydroxynonenal (4-HNE) [3]. These aldehydes are highly reactivebut considerably more selective than the parent ROS. Specificmodification of proteins by 4-HNE and similar aldehydesoften affects protein function. This is probably the basis ofthe signalling role of 4-HNE at physiological concentrations[4,5] and the toxicity elicited by this compound at highconcentrations [3].The generation of 4-HNE, primarily from arachidonic acid [3],and the formation of 4-HNE–protein adducts [6] are wellestablishedphenomena in mammalian tissues. Although thefruitfly, Drosophila melanogaster, lacks arachidonic acid [7,8], 4-HNE can be formed from any ω-6 polyunsaturated fatty acid [9].In addition to arachidonic acid, this also includes linoleic acid,which is present in large amounts in larval and adult stages ofDrosophila [8,10]. Although the ability of Drosophila to desaturatefatty acids is limited, dietary polyunsaturated fatty acids arereadily incorporated into phospholipids [8]. Therefore 4-HNEandor similar 4-hydroxyalkenals are expected to be generated inDrosophila. We have previously confirmed the presence of 4-HNE–protein adducts in Drosophila by immunoblotting [11],and Yan and Sohal [12] have demonstrated the formation ofsuch adducts in the housefly.In mammals, the major route of 4-HNE metabolism is viaglutathione conjugation [13], catalysed by a specialized subclassof Alpha-class glutathione S-transferases (GSTs) [14,15], althougha reductive pathway may be significant in some tissues.Because of the apparently universal and obligatory formation of4-hydroxyalkenals in aerobes, we examined which of the DrosophilaGSTs has the ability to conjugate 4-HNE. In a previousstudy [11], we investigated DmGSTS1-1 (see the Materials andmethods section for an explanation of the nomenclature), whichwas previously assumed to be a structural muscle protein orperhaps a stretch sensor devoid of enzymic properties [16].However, DmGSTS1-1 has moderate glutathione-conjugatingactivity for 4-HNE, and due to its high level of expression, itaccounts for most of this activity in adult Drosophila [11]. In thecourse of that study, we also found a less abundant GST (ormixture of GSTs), probably belonging to the Delta class, whichhad a higher specific activity for 4-HNE conjugation thanDmGSTS1-1 [11]. We have now characterized a number ofDrosophila GSTs, both of the Delta and the novel Epsilon class,and have shown that some, but not all, of them participate in4-HNE metabolism.Abbreviations used: CDNB, 1-chloro-2,4-dinitrobenzene; EST, expressed sequence tag; GST, glutathione S-transferase; 4-HNE, 4-hydroxynonenal;ORF, open reading frame; poly(A)+, polyadenylated; ROS, reactive oxygen species; RT, reverse transcriptase.1 To whom correspondence should be addressed (e-mail zimniakpiotruams.edu). 2003 Biochemical Society

662 R. Sawicki and othersMATERIALS AND METHODSNomenclature of Drosophila GSTsIn the present study, we adhere to a recently proposed unifiednomenclature that includes insect as well as mammalian GSTs[17]. Accordingly, we use the term ‘Delta class’ for DrosophilaGST-1 and its homologues [18], refer to the protein originallydescribed as GST-2 [19] as DmGSTS1-1 and designate DrosophilaGST-3 [20] as DmGSTE1-1, a prototypical member of theEpsilon class of GSTs (Professor Philip G. Board, personal communication).This nomenclature has been recently adopted byFlybase (www.flybase.org).Cloning of Drosophila Delta- and Epsilon-class GSTsPCR primers specific for the ten Delta-class genes and forDmGSTE1 were used to amplify the sequences from Drosophilagenomic DNA. The primers are listed in Table 1. All upstreamprimers were designed to span the translation initiation codonand to introduce an NdeI restriction site to facilitate subsequentsubcloning. For the same reason, the downstream primersintroduced a unique restriction site, as specified in Table 1. Onehundred male adult flies were homogenized, and total genomicDNA was isolated using the DNeasy Tissue Kit (Qiagen). EachPCR reaction contained, in a total volume of 25 µl, 50 ng ofDrosophila genomic DNA as the template and 25 ng each of therespective primers. Reactions were initiated by the addition ofTaq polymerase using the hot start technique and were continuedfor 32 cycles of denaturation for 30 s at 95 C, annealing for 30 sat 48 C for DmGSTD6, 50C for DmGSTD10, ramped from52.2 to 49 C over 32 cycles (touch-down PCR) for DmGSTD7and from 56.4 to 50 C for DmGSTD1–5, DmGSTD8, DmGSTD9and DmGSTE1, followed by extension for 1 min at 72 Ccycle.The specificity of the amplification was verified by sequencing ofall PCR products.Bacterial expression and purification of Drosophila GSTsThe PCR-amplified fragments encoding the Drosophila GSTswere digested with NdeI and a restriction enzyme appropriate forthe antisense primer (see Table 1) and subcloned between thecorresponding sites of pET-30a() (Novagen). Escherichia coliBL21(DE3)pLysS cells were used for expression of all proteinsexcept for DmGSTD3-3, DmGSTD7-7 and DmGSTE1-1, whichwere expressed in the BL21 Star(DE3)pLysS strain, which carriesa truncated RNAse E, leading to increased mRNA stability.Expression was induced with 1 mM isopropyl β-D-thiogalactosideat absorbance A 0.4–0.6 and it was continued for 8 h at37 C. The bacteria were collected by centrifugation and thepellets were frozen at 20 C until use. Bacterial cells were lysedand the GSTs, except for DmGSTD3-3 and DmGSTE1-1, werepurified by glutathione affinity chromatography using a 1 mlprepacked glutathione Sepharose 4 Fast Flow column(Amersham) according to the manufacturer’s instructions.DmGSTD3-3 and DmGSTE1-1 did not bind to the glutathionecolumn. The bacterial lysate (from 100 ml culture) containingDmGSTD3-3 was fractionated by ammonium sulphate precipitation.The 25–50% fraction was dialysed against 10 mMpotassium phosphate (pH 7.4) and further fractionated on aSephadex G-200 column (1 cm50 cm). Fractions containingthe GST were pooled and loaded on a hydroxyapatite column(1 cm8.5 cm). The column was eluted with a gradient of10–400 mM potassium phosphate (pH 6.5). Fractions containingDmGSTD3-3 eluted at 300 mM KCl and were identified bySDSPAGE. DmGSTE1-1 was purified by ion-exchange chromatography.Frozen bacterial cells from 500 ml of culture werelysed by sonication in 20 mM TrisHCl (pH 7.5), 50 mM KCl,centrifuged for 30 min at 30000 g and the supernatant waspassed over a DEAE column (Macro-Prep DEAE Support; Bio-Rad Laboratories), which did not retain DmGSTE1-1. The flowthroughwas then applied to a Macro-Prep High Q Supportcolumn (0.5 cm3 cm; Bio-Rad Laboratories), which waswashed with 20 mM TrisHCl (pH 7.5), 50 mM KCl and elutedwith a step gradient of KCl in 20 mM TrisHCl (pH 7.5). The4-HNE-conjugating activity eluted at 0.1 M KCl. The fractionwas dialysed against 50 mM TrisHCl (pH 7.5) and 1.3 mM2-mercaptoethanol.Determination of enzyme activityGST activities were measured spectrophotometrically in a microtitreplate reader (SpectraMax Plus; Molecular Devices, Sunnyvale,CA, U.S.A.). Conjugation of 1-chloro-2,4-dinitrobenzene(CDNB) was measured at 25 C. Enzyme activity with 4-HNEwas calculated from the rate of consumption of this substrate at30 C [13]. To determine kinetic constants of 4-HNE conjugation,the GST activities of the enzymes were measured by varying theconcentrations of 4-HNE at fixed concentrations of reducedglutathione. The results were analysed by least-squares nonlinearfitting of a Michaelis–Menten hyperbola or the HillTable 1Primers used to amplify D. melanogaster GST genesForward primers were designed to introduce an NdeI restriction site (underline) which includes the initiation codon ATG. Reverse primers spanned all or part of the stop codon (double underline)and were designed to introduce a unique restriction site (lower-case boldface; EcoRI for DmGSTD1, DmGSTD3, DmGSTD5, DmGSTD7–9 and DmGSTE1; BamHI for DmGSTD2 and DmGSTD10;KpnI for DmGSTD4; SacI for DmGSTD6).Clone Forward primer Reverse primerGSTD1 5-CTTCTACAGTCATATGGTTGACTTCTACTAC 5-AAAACGTgaattcCAGGCTTATTCGSTD2 5-GCAATATCCCCCATATGGACTTTTACTAC 5-ATAGTTGggatccCGATCACTTAGCGSTD3 5-CGCTCCGTTCATATGGTGGGCAAG 5-TCgaattcAATGAATTATTTAGCAGCATTCTGGSTD4 5-CGCCCCAGCCATATGGATTTCTACT 5-GTTTggtaccACTGTAGTTACTTTAATGAGGSTD5 5-ACACTCAAATAACCCATATGGATTTCTATT 5-TTgaattcTTATTGCTTCGCCGCCGSTD6 5-GCCCTTCTGCATATGGATCTCTATA 5-CGgagctcTCTATAATATTTATAACTTTTCGSTD7 5-GTGAGTTCTACATATGACAAACATATTC 5-GTgaattcAAAACTATTCTACGGTGGSTD8 5-GGACCTTTCCATATGGACTTTTACTACC 5-CATTTACTGCAGgaattcTCATCAAATCGSTD9 5-CAGCGAGTACATATGTTGGACTTCTAC 5-GATgaattcTTACAGACCCGTCATTGGSTD10 5-GAGATAGCGTCATATGGATTTATACTATAG 5-TCGggatccTATATTCAGTTATCGAAGGSTE1 5-ACAACACATCATATGTCGAGCTCTGG 5-TTgaattcACTTGTCGCCCAGC 2003 Biochemical Society

Cloning and expression of Drosophila glutathione S-transferases663equation. Glutathione peroxidase activity with cumene hydroperoxideas substrate was determined as described in [21].Reverse transcriptase (RT)–PCRThe presence of DmGSTD3 transcripts was determined in totalRNA isolated from wandering third-instar Drosophila larvaeusing the RNeasy Mini Kit (Qiagen). The RNA was treated withRNase-free DNase I (Roche Applied Science, Indianapolis, IN,U.S.A.; 1 unit of DNase Iµg of RNA in 25 mM TrisHCl,5 mM MgCl , 0.1 mM EDTA, at 37 C for 30 min, followed byheat inactivation of DNase at 75 C for 15 min). RNA (200 ng)was used for each RT–PCR reaction (Titan One Tube RT–PCRSystem; Roche Applied Science). The sense primer was 5-CGCTCCGTTCTGATGGTGGGCAAG, and the antisenseprimer was 5-TCGTTATAAATGAATTATTTAGCAGCAT-TCTG. The reverse transcription step was performed at 50 Cfor 30 min, followed by inactivation of the RT at 94 C for 2 min.The conditions for the subsequent PCR step were as follows: 10cycles of 94 C, 10 s; 58 C, 30 s; 68 C, 45 s and then 25 cyclesof 94 C, 10 s; 58 C, 30 s; 68 C, initially 45 s with cycleelongation by 5 s for each cycle; 1 cycle at 94 C, 10 s; 58 C,30 s; 68 C for 7 min. In control reactions, the mixture of RT andthermostable DNA polymerase supplied in the Titan One TubeRT–PCR System was substituted with Taq polymerase (Biolase;Bioline USA, Canton, MA, U.S.A.). For detection of theDmGSTD7 transcript, polyadenylated [poly(A)+] RNA was isolatedfrom 25 µg of total larval RNA using the Oligotex DirectmRNA Mini Kit (Qiagen), and one-tenth of each of the recoveredpoly(A)+ RNA was used for every RT–PCR or control reactionunder conditions described above for DmGSTD3, with the senseprimer 5-GTGAGTTCTAACAATGACAAACATATTC andthe antisense primer 5-GTCAAACGAAAACCATTCTACG-GTG.RESULTSIdentification of gene clusters encoding Delta- and Epsilon-classGSTs in Drosophila genomeThe genomic sequence of D. melanogaster was examined, usingthe BLAST software, for the presence of open reading frames(ORFs) homologous with previously described Delta [18] andEpsilon [20] GSTs. The Delta cluster at chromosomal position87B11–87B13 spanned approx. 18 kb and contained the eightgenes previously designated as D1, D21, D22, D23, D24, D25,D26 and D27 [18,22,23]. In keeping with a recent nomenclatureproposal [17], we refer to these genes as DmGSTD1 to DmGSTD8respectively. In addition, the Delta cluster contained two recentlyreported [17] but not further characterized genes DmGSTD9and DmGSTD10. Three genes (DmGSTD1, DmGSTD9 andDmGSTD10) were located on the same DNA strand, whereas theremaining Delta-class genes were on the opposite strand. Theorientation and relative positions of the Delta ORFs are depictedin Figure 1(A).A novel Theta class-related GST gene Gst-3 has been describedrecently [20]. According to a suggestion from Professor PhilipG. Board (personal communication), we refer to this gene asDmGSTE1, a member of the Epsilon class of GSTs. By BLASTsearches of the Drosophila genomic sequence, we identified nineadditional genes (DmGSTE2–DmGSTE10) with high homologyto DmGSTE1. The Epsilon genes form a tight cluster spanningapprox. 13 kb at chromosomal position 55C9. Within that cluster(Figure 1B), one gene (DmGSTE10) has an antiparallel orientationrelative to the remaining genes.Figure 1 Clusters of Delta- and Epsilon-class glutathione transferasegenes in D. melanogasterThe map is drawn to scale and is based on the Drosophila genomic sequence. The arrowsdenote the positions and 5 3 direction of the coding sequences of the genes; none of theDelta and Epsilon genes contains introns within the ORF. The chromosomal positions of theclusters are indicated.Table 2 Context of the initiation codon of D. melanogaster Delta andEpsilon GST genesGene 4 3 2 1 Start 4DmGSTD1 T A A A ATG GDmGSTD2 C A A C ATG GDmGSTD3 T C T G ATG GDmGSTD4 C A A C ATG GDmGSTD5 C G A A ATG GDmGSTD6 G A C G ATG GDmGSTD7 A A C A ATG ADmGSTD8 C A T C ATG GDmGSTD9 A A T C ATG TDmGSTD10 T A A G ATG GDmGSTE1 T A T C ATG TDmGSTE2 C A T C ATG TDmGSTE3 A G A C ATG GDmGSTE4 G A G A ATG GDmGSTE5 A A A C ATG GDmGSTE6 C A A G ATG GDmGSTE7 C A A G ATG CDmGSTE8 C G G C ATG TDmGSTE9 A G C G ATG GDmGSTE10 C A C A ATG GDrosophilaConsensus [31] 3% G 13% G 9% G 18% G ATG 26% G32% A 82% A 56% A 38% A ATG 37% A8% T 1% T 10% T 8% T ATG 22% T57% C 3% C 25% C 36% C ATG 15% CSequence similarities within the Delta and Epsilon clusters ofGSTsIt has been previously reported that the coding sequences of theDelta-class genes [18,24,25] as well as of the DmGSTE1 gene [20]contain no introns. We have now extended this observation toDmGSTD9, DmGSTD10 and the nine novel DmGSTE genes. Allof the Delta and Epsilon genes can be conceptually translated.Their initiation codons are in a context appropriate for Drosophila[26] (Table 2) and they are followed by an ORF of a sizeexpected for glutathione transferases (see the Discussion sectionfor additional details). The alignment of the resulting proteinsequences is shown in Figure 2. The Delta-class sequences formeda tight cluster, with the exception of DmGSTD7-7, which carrieda 47-amino-acid N-terminal extension and was truncated at theC-terminus by approx. 50 amino acids, and of DmGSTD3-3,which was shorter by approx. 15 residues at the N-terminus,as compared with the remaining Delta-class proteins. The 2003 Biochemical Society

664 R. Sawicki and othersFigure 2Alignment of the predicted protein sequences of D. melanogaster Delta- and Epsilon-class glutathione transferasesResidues identical in at least 18 of the 20 sequences are shaded in black, and residues similar (as defined by the BLOSUM62 matrix) in at least 18 of the 20 sequences are shown as blackletters shaded in light grey. Residues that are, by the above criteria, neither identical nor similar in the entire set of 20 sequences but are similar (or identical) in at least 9 out of 10 sequencesin one of the two GST clusters are shown as white letters shaded in dark grey. Thus amino acid positions denoted as white letters shaded in dark grey are conserved within the Delta or the Epsiloncluster, but not between the two clusters. 2003 Biochemical Society

Cloning and expression of Drosophila glutathione S-transferases665Figure 3 Percentage of identity/similarity respectively for pairwise alignmentsof Drosophila Delta-class protein sequencesThe PAM250 matrix was used to calculate similarity scores.similarities and identities of all pairs of the Delta-class sequencesare presented in Figure 3. Similarly, all Epsilon-class proteinscould be brought into close alignment, with DmGSTE10-10 beingthe most divergent sequence because of a C-terminal extensionof approx. 15 amino acids (Figure 2). The similarity between theDelta and Epsilon clusters was less pronounced. For example,the similarity and identity between the DmGSTD1-1 andDmGSTE1-1 proteins was only 36 and 49% respectively. Therelationship between the two GST classes can be visualized as ahypothetical phylogenetic tree, which shows that the Delta andEpsilon nucleotide sequences form distinct clusters, with bothclusters only distantly related to the remaining characterizedDrosophila GST gene, DmGSTS1 (Figure 4). The sequencehomology within each of the two clusters, together with thephysical proximity of all Delta genes on chromosome 3 and allEpsilon genes on chromosome 2 (Figure 1), suggests that eachcluster was probably formed by repeated duplication events of aDelta and an Epsilon ancestral gene respectively. Interestingly, inthe Epsilon cluster, similarities in sequence and thus probableevolutionary relationships of individual genes are reflected intheir physical proximity. For example, the DmGSTE5 andDmGSTE6 genes, which have probably diverged most recentlyon an evolutionary time scale (Figure 4), are located next to eachother on the genomic DNA (Figure 1). This generalization holdsfor all members of the Epsilon group of genes, but not for theDelta cluster. This indicates that, in the Epsilon cluster, individualgenes (rather than blocks of genes) underwent duplication, withno subsequent rearrangement.Cloning and bacterial expression of Delta- and Epsilon-class GSTsThe ten Delta-class genes (DmGSTD1–DmGSTD10) as well asDmGSTE1 were amplified by PCR using genomic DNA as thetemplate, as described in Materials and methods section. Theirintron-free coding sequences were used directly for expression ofthe proteins in E. coli. All proteins could be expressed, albeit withdifferent yields. Although the expression level of DmGSTD5-5and DmGSTD7-7 was low (results not shown), it was sufficientto determine the enzymic activity of the proteins in bacteriallysates. Enzymes selected because of their ability to conjugate4-HNE (see below) were further purified (Figure 5).Figure 4Phylogenetic tree of D. melanogaster GSTsThe tree representing similarities among the full-length protein-encoding nucleotide sequenceswas calculated using the KITSCH program [49], as implemented in the BioEdit program suite(www.mbio.ncsu.edu/BioEdit). The program assumes a constant evolutionary rate; the barindicates the time in which 0.1 nucleotide substitution occurs at any given site.Catalytic activity of Drosophila GSTsExcept for DmGSTD3-3, all Delta-class GSTs conferred CDNBconjugatingactivity on lysates of bacterial cells in which theywere expressed. In contrast, DmGSTD3-3 and DmGSTE1-1 hadno activity with CDNB but were able to conjugate 4-HNE incrude bacterial lysates. This pattern of activities resembles that ofthe quantitatively predominant Drosophila GST, DmGSTS1-1([11] and Table 3). Six of the 10 Delta-class GSTs had 4-HNEconjugatingactivity. Thus each of the 11 enzymes under studywas active with either 4-HNE or CDNB. Enzymes able tocatalyse the conjugation of 4-HNE were of special interest in thecontext of the present study and were selected for purificationand further characterization. The kinetic parameters of the latterGSTs are shown in Table 3.In this group, DmGSTD1-1 had thehighest catalytic efficiency for 4-HNE conjugation, only 4-foldlower than the catalytic efficiency of the highly specialized murineenzyme mGSTA4-4 [21]. DmGSTD7-7 had an appreciablecatalytic efficiency for 4-HNE which exceeded that of DmGSTS1-1. The remaining four Delta-class GSTs had lower but clearlymeasurable activities for 4-HNE. As observed by us previously[11], DmGSTS1-1 contributes most of 4-HNE-conjugating activityof adult Drosophila because of the high abundance of theprotein. Our present results indicate that DmGSTD1-1 andseveral additional GSTs have higher or equal catalytic-centreactivities andor catalytic efficiencies compared with DmGSTS1- 2003 Biochemical Society

Cloning and expression of Drosophila glutathione S-transferases667Figure 6 Identification of DmGSTD3 and DmGSTD7 transcripts in Drosophilalarvae by RT–PCRTotal RNA was used as the starting material for amplification of DmGSTD3, and poly(A)+ RNAwas used to amplify DmGSTD7. The identity of the amplification products (denoted by arrows)was confirmed by partial sequencing; the additional bands were probably the result ofmispriming and were not further analysed. Control reactions lacking RT (lanes labelled RT)did not yield amplification products.(results not shown). Therefore the poly(A)+ RNA fraction waspurified from total RNA for amplification of the latter sequence.The results of RT–PCR demonstrate that the DmGSTD3 andDmGSTD7 genes are expressed in Drosophila larvae, althoughthe level of the DmGSTD7 mRNA may be lower than that ofDmGSTD3.DISCUSSIONThe cluster of D. melanogaster Delta-class GSTs has beenoriginally described, and several of the enzymes have beenbiochemically characterized in a series of elegant papers byC.-P. D. Tu et al. [18,22–24,28,29]. Two additional membersof the cluster have been identified recently by bioinformaticsapproaches [17] but have not been biochemically characterized.Similarly, the DmGSTE1 (Gst-3) gene has been reported [20], butthe functional properties of the protein remained unknown. Wehave now cloned, expressed in E. coli and characterized all tenDelta-class GSTs and DmGSTE1-1. Furthermore, we identified,but have not further characterized, nine additional members ofthe Epsilon cluster. Thus both the Delta- and the Epsilon-classGSTs form families of at least ten members each, probably as theresult of repeated gene duplications.DmGSTE1-1 and all ten Delta-class GSTs could be expressedin E. coli as enzymically active proteins. This result appears tocontradict the earlier conclusion that DmGSTD3 and DmGSTD7are likely to be pseudogenes [18]. DmGSTD3 was presumed to bea pseudogene on the basis of a sub-optimal Kozak context of thetranslation-initiation codon, which may prevent its synthesis,and a 15-amino-acid N-terminal truncation, which may renderthe protein inactive even if synthesized [18]. Our demonstrationof bacterial expression of DmGSTD3-3 does not address thequestion of whether mGSTD3 mRNA can be efficiently translatedin eukaryotic, and specifically Drosophila, cells. However, thedatabase of all known and predicted Drosophila transcripts ([30];file available as ‘nagadfly.dros.RELEASE2’ on www.fruitfly.org) contains, in addition to DmGSTD3, at least twoother sequences that have an identical context of the initiatorATG (TCTG ATG G) and encode identifiable proteins: a Zndependentexopeptidase (Flybase symbol CG10073) and a celladhesion protein (CG5550). This is consistent with the fact thatthe Kozak consensus is not an absolute requirement for translation,but rather one of several interdependent factors that codetermineinitiation efficiency. Moreover, Drosophila appears totolerate more deviation from the ‘ideal’ context of the initiatorcodon than mammalian cells [26,31]. In fact, suboptimal determinantsof protein synthesis initiation may render the processmore amenable to regulation [32], an aspect that may be relevantfor some of the GSTs (see below). An expressed sequence tag(EST) clone (LP11313, GenBank accession nos. AI297103 andAY118350) has been isolated that fully matches the genomicDmGSTD3 sequence, except for the presence of a poly(A)+tail at the 3-end of the EST clone. Finally, we identified aDmGSTD3 transcript by RT–PCR. This demonstrates thatthe DmGSTD3 gene is expressed, at least up to the stage oftranscription and transcript processing. This argues against thepossibility that DmGSTD3 is a pseudogene.As noted previously [18], DmGSTD3-3 lacks the N-terminalsequence (cf. Figure 2) that includes the highly conserved tyrosineresidue involved in the catalytic cycle of GSTs. Strikingly,DmGSTD3-3 is active towards 4-HNE, with kinetic parameterssimilar to those of other Delta-class GSTs (Table 3). Thisindicates that other residues may take the place of the active-sitetyrosine.DmGSTD7 was deemed to be a pseudogene because of a C-terminal truncation which may affect the electrophile-bindingpocket of the active site andor protein stability [18]. We havedemonstrated by RT–PCR the presence of DmGSTD7 mRNA inDrosophila larvae. Furthermore, our results show that theDmGSTD7-7 protein is active towards both 4-HNE and CDNB(Table 3). In fact, its catalytic efficiency for 4-HNE was secondhighest among the enzymes tested in the present study. Theexamples of DmGSTD3-3 and DmGSTD7-7 show that GSTscan tolerate significant deletions, probably through compensatorychanges in the folding of the polypeptide.It is noteworthy that six out of the ten Delta-class GSTs, aswell as DmGSTE1-1, had significant activity for 4-HNE conjugation.Together with the previously characterized DmGSTS1-1 [11], there are at least seven distinct GSTs in Drosophila withthis activity. The catalytic efficiency of one of the Delta-classenzymes, DmGSTD1-1, was lower by only a factor of 4 whencompared with that of mammalian Alpha-class GSTs highlyspecialized for 4-HNE conjugation, exemplified by mGSTA4-4[21]. At least three of the enzymes, namely DmGSTD3-3,DmGSTE1-1 (Table 3) and DmGSTS1-1 [11], are active towards4-HNE but not towards the near-universal GST model substrateCDNB. Finally, 4-HNE-conjugating activity probably arose onmultiple independent occasions during evolution, since it isassociated with some, but not all, of the Theta-related enzymesin Drosophila [DmGSTS1-1 [11] and certain Delta- and EpsilonclassGSTs (this work)], with some, but again not all, Alpha-classmammalian GSTs [15,33] and with at least one Pi-class GST inCaenorhabditis elegans [34]. Collectively, these findings suggestthat conjugation of 4-HNE evolved, and is maintained, byselective pressure, rather than being an adventitious activity ofsome GSTs.In addition to toxicity at supraphysiological levels, 4-HNE hassignalling functions affecting, among others, cell proliferation,differentiation and apoptosis (see e.g. [35–39]). It is tempting tospeculate that the multiplicity of GSTs involved in 4-HNE 2003 Biochemical Society

668 R. Sawicki and othersmetabolism is necessary to modulate and terminate thesefunctions of 4-HNE in a variety of physiological situations,perhaps including embryogenesis and insect metamorphosis. Thelack of DmGSTD3, DmGSTD5, DmGSTD7 and DmGSTD8transcripts detectable by Northern blotting in adult Drosophila[18] may indicate that their expression is spatially and temporallyrestricted, perhaps to pre-adult stages. For example, the ESTrepresenting DmGSTD3 (LP11313; Berkeley DrosophilaGenome Project, direct submission to GenBank) has beenisolated from a larvalearly pupal stage of Drosophila, and wehave demonstrated the presence of DmGSTD3 and DmGSTD7transcripts in third instar larvae. In addition to transcriptionalregulation, the suboptimal Kozak context of the translationalstart site of some of the above GSTs is consistent with translationalregulation. Low and tightly controlled expression ischaracteristic of regulatory proteins, but would be unusual for anenzyme whose primary function is general detoxification.Previously, insect GSTs have been studied primarily becauseof their possible involvement in the metabolic inactivation ofinsecticides [40]. Our work indicates that several DrosophilaGSTs participate in defence mechanisms against oxidative stress,and in the metabolism of endogenously formed lipid peroxidationproducts. In this, they functionally resemble mammalian AlphaclassGSTs [41,42]. A possible role of lipid peroxidation products,and thus GSTs that metabolize them, in the modulation ofembryogenesis and development has been already discussed.Another process in which oxidative stress plays a key role isorganismal aging [43,44]. In this context, it is intriguing thatlifespan in Drosophila correlates with the expression level ofDmGSTS1-1 [45] and DmGSTD1-1 [46], two enzymes identifiedby us as capable of metabolizing 4-HNE.The above examples demonstrate that the physiological rolesof GSTs transcend simple detoxification and affect a number offundamental biological processes. Drosophila is well suited to thestudy of these processes because of its known genetics andavailability of methods for molecular interventions. Since basicbiological processes are often conserved between invertebratessuch as Drosophila and mammals, including humans, we expectthat the elucidation of the physiological functions of DrosophilaGSTs will have broad biological implications.We thank Mr Venu Varma for computational help with the identification, in thedatabase of genes potentially expressed in Drosophila, of transcripts with asuboptimal Kozak context which corresponds to that of DmGSTD3. This investigationwas supported in part by the National Institute of Aging (USPHS grant no.AG018845) and the National Institute of Environmental Health Sciences (grantno. ES07804).REFERENCES1 Halliwell, B. and Gutteridge, J. M. C. (1999) Free Radicals in Biology and Medicine,Oxford University Press, Oxford2 Schneider, C., Tallman, K. A., Porter, N. A. and Brash, A. R. (2001) Two distinctpathways of formation of 4-hydroxynonenal: mechanisms of nonenzymatictransformation of the 9- and 13-hydroperoxides of linoleic acid to 4-hydroxyalkenals.J. Biol. Chem. 276, 20831–208383 Esterbauer, H. (1993) Cytotoxicity and genotoxicity of lipid-oxidation products. Am. J.Clin. Nutr. 57, 779S–786S4 Dianzani, M. U. (1998) 4-Hydroxynonenal and cell signalling. Free Radic. Res. 28,553–5605 Uchida, K., Shiraishi, M., Naito, Y., Torii, Y., Nakamura, Y. and Osawa, T. (1999)Activation of stress signaling pathways by the end product of lipid peroxidation.4-Hydroxy-2-nonenal is a potential inducer of intracellular peroxide production.J. Biol. Chem. 274, 2234–22426 Uchida, K. and Stadtman, E. R. (1992) Modification of histidine residues in proteinsby reaction with 4-hydroxynonenal. Proc. Natl. Acad. Sci. U.S.A. 89, 4544–45487 Stark, W. S., Lin, T. N., Brackhahn, D., Christianson, J. S. and Sun, G. Y. (1993)Fatty acids in the lipids of Drosophila heads: effects of visual mutants, carotenoiddeprivation and dietary fatty acids. Lipids 28, 345–3508 Draper, H. H., Philbrick, D. P., Agarval, S., Meidiger, R. and Phillips, J. P. (2000)Avid uptake of linoleic acid and vitamin E by Drosophila melanogaster. Nutr. Res. 20,113–1209 Pryor, W. A. and Porter, N. A. (1990) Suggested mechanisms for the production of4-hydroxy-2-nonenal from the autoxidation of polyunsaturated fatty acids. Free Radic.Biol. Med. 8, 541–54310 Jones, H. E., Harwood, J. L., Bowen, I. D. and Griffiths, G. (1992) Lipid compositionof subcellular membranes from larvae and prepupae of Drosophila melanogaster.Lipids 27, 984–98711 Singh, S. P., Coronella, J. A., Benes , H., Cochrane, B. J. and Zimniak, P. (2001)Catalytic function of Drosophila melanogaster glutathione S-transferase DmGSTS1-1(GST-2) in conjugation of lipid peroxidation end products. Eur. J. Biochem. 268,2912–292312 Yan, L. J. and Sohal, R. S. (1998) Mitochondrial adenine nucleotide translocase ismodified oxidatively during aging. Proc. Natl. Acad. Sci. U.S.A. 95, 12896–1290113 Alin, P., Danielson, U. H. and Mannervik, B. (1985) 4-Hydroxyalk-2-enals aresubstrates for glutathione transferase. FEBS Lett. 179, 267–27014 Awasthi, Y. C., Zimniak, P., Awasthi, S., Singhal, S. S., Srivastava, S. K., Piper, J. T.,Chaubey, M., Petersen, D. R., He, N.-G., Sharma, R. et al. (1996) A new group ofglutathione S-transferases with protective role against lipid peroxidation. InGlutathione S-transferases: Structure, Function and Clinical Implications (Vermeulen,N. P. E., Mulder, G. J., Nieuwenhuyse, H., Peters, W. H. M. and van Bladeren, P. J.,eds.), pp. 111–124, Taylor & Francis, London15 Hubatsch, I., Ridderstrom, M. and Mannervik, B. (1998) Human glutathionetransferase A4-4: an Alpha class enzyme with high catalytic efficiency in theconjugation of 4-hydroxynonenal and other genotoxic products of lipid peroxidation.Biochem. J. 330, 175–17916 Clayton, J. D., Cripps, R. M., Sparrow, J. C. and Bullard, B. (1998) Interaction oftroponin-H and glutathione S-transferase-2 in the indirect flight muscles of Drosophilamelanogaster. J. Muscle Res. Cell Motil. 19, 117–12717 Chelvanayagam, G., Parker, M. W. and Board, P. G. (2001) Fly fishing for GSTs: aunified nomenclature for mammalian and insect glutathione transferases. Chem. Biol.Interact. 133, 256–26018 Toung, Y.-P. S., Hsieh, T. and Tu, C.-P. D. (1993) The glutathione S-transferase Dgenes. A divergently organized, intronless gene family in Drosophila melanogaster.J. Biol. Chem. 268, 9737–974619 Beall, C., Fyrberg, C., Song, S. and Fyrberg, E. (1992) Isolation of a Drosophila geneencoding glutathione S-transferase. Biochem. Genet. 30, 515–52720 Singh, M., Silva, E., Schulze, S., Sinclair, D. A. R., Fitzpatrick, K. A. and Honda,B. M. (2000) Cloning and characterization of a new theta-class glutathione S-transferase (GST) gene, gst-3, from Drosophila melanogaster. Gene 247, 167–17321 Zimniak, P., Singhal, S. S., Srivastava, S. K., Awasthi, S., Sharma, R., Hayden, J. B.and Awasthi, Y. C. (1994) Estimation of genomic complexity, heterologous expression,and enzymatic characterization of mouse glutathione S-transferase mGSTA4-4 (GST5.7). J. Biol. Chem. 269, 992–100022 Tang, A. H. and Tu, C.-P. D. (1994) Biochemical characterization of Drosophilaglutathione S-transferases D1 and D21. J. Biol. Chem. 269, 27876–2788423 Lee, H. C. and Tu, C.-P. D. (1995) Drosophila glutathione S-transferase D27:functional analysis of two consecutive tyrosines near the N-terminus. Biochem.Biophys. Res. Commun. 209, 327–33424 Toung, Y. P., Hsieh, T. S. and Tu, C.-P. D. (1991) The Drosophila glutathioneS-transferase 1-1 is encoded by an intronless gene at 87B. Biochem. Biophys.Res. Commun. 178, 1205–121125 Lougarre, A., Bride, J. M. and Fournier, D. (1999) Is the insect glutathioneS-transferase I gene family intronless? Insect Mol. Biol. 8, 141–14326 Feng, Y., Gunter, L. E., Organ, E. L. and Cavener, D. R. (1991) Translation initiation inDrosophila melanogaster is reduced by mutations upstream of the AUG initiatorcodon. Mol. Cell. Biol. 11, 2149–215327 Xiao, B., Singh, S. P., Nanduri, B., Awasthi, Y. C., Zimniak, P. and Ji, X. (1999)Crystal structure of a murine glutathione S-transferase in complex with a glutathioneconjugate of 4-hydroxynon-2-enal in one subunit and glutathione in the other:evidence of signaling across the dimer interface. Biochemistry 38, 11887–1189428 Toung, Y. P. and Tu, C.-P. D. (1992) Drosophila glutathione S-transferases havesequence homology to the stringent starvation protein of Escherichia coli. Biochem.Biophys. Res. Commun. 182, 355–36029 Tang, A. H. and Tu, C.-P. D. (1995) Pentobarbital-induced changes in Drosophilaglutathione S-transferase D21 mRNA stability. J. Biol. Chem. 270, 13819–1382530 Rubin, G. M., Yandell, M. D., Wortman, J. R., Gabor Miklos, G. L., Nelson, C. R.,Hariharan, I. K., Fortini, M. E., Li, P. W., Apweiler, R., Fleischmann, W. et al. (2000)Comparative genomics of the eukaryotes. Science 287, 2204–221531 Cavener, D. R. (1987) Comparison of the consensus sequence flanking translationalstart sites in Drosophila and vertebrates. Nucleic Acids Res. 15, 1353–136132 Kozak, M. (1999) Initiation of translation in prokaryotes and eukaryotes. Gene 234,187–208 2003 Biochemical Society

Cloning and expression of Drosophila glutathione S-transferases66933 Cheng, J. Z., Yang, Y., Singh, S. P., Singhal, S. S., Awasthi, S., Pan, S. S.,Singh, S. V., Zimniak, P. and Awasthi, Y. C. (2001) Two distinct 4-hydroxynonenalmetabolizing glutathione S-transferase isozymes are differentially expressed in humantissues. Biochem. Biophys. Res. Commun. 282, 1268–127434 Engle, M. R., Singh, S. P., Nanduri, B., Ji, X. and Zimniak, P. (2001) Invertebrateglutathione transferases conjugating 4-hydroxynonenal: CeGST 5.4 fromCaenorhabditis elegans. Chem. Biol. Interact. 133, 244–24835 Barrera, G., Pizzimenti, S., Muraca, R., Barbiero, G., Bonelli, G., Baccino, F. M.,Fazio, V. M. and Dianzani, M. U. (1996) Effect of 4-hydroxynonenal on cell cycleprogression and expression of differentiation-associated antigens in HL-60 cells. FreeRadicals Biol. Med. 20, 455–46236 Liu, W., Akhand, A. A., Kato, M., Yokoyama, I., Miyata, T., Kurokawa, K., Uchida, K.and Nakashima, I. (1999) 4-Hydroxynonenal triggers an epidermal growth factorreceptor-linked signal pathway for growth inhibition. J. Cell Sci. 112, 2409–241737 Pizzimenti, S., Barrera, G., Dianzani, M. U. and Brusselbach, S. (1999) Inhibition ofD1, D2, and A-cyclin expression in HL-60 cells by the lipid peroxidation product4-hydroxynonenal. Free Radicals Biol. Med. 26, 1578–158638 Cheng, J.-Z., Singhal, S. S., Saini, M., Singhal, J., Piper, J. T., Van Kuijk, F. J. G. M.,Zimniak, P., Awasthi, Y. C. and Awasthi, S. (1999) Effects of mGST A4 transfectionon 4-hydroxynonenal-mediated apoptosis and differentiation of K562 humanerythroleukemia cells. Arch. Biochem. Biophys. 372, 29–3639 Rinaldi, M., Barrera, G., Spinsanti, P., Pizzimenti, S., Ciafre, S. A., Parella, P., Farace,M. G., Signori, E., Dianzani, M. U. and Fazio, V. M. (2001) Growth inhibition anddifferentiation induction in murine erythroleukemia cells by 4-hydroxynonenal. FreeRadical Res. 34, 629–63740 Ranson, H., Rossiter, L., Ortelli, F., Jensen, B., Wang, X. L., Roth, C. W.,Collins, F. H. and Hemingway, J. (2001) Identification of a novel class ofinsect glutathione S-transferases involved in resistance to DDT in the malariavector Anopheles gambiae. Biochem. J. 359, 295–30441 Cheng, J.-Z., Sharma, R., Yang, Y., Singhal, S. S., Sharma, A., Saini, M. K.,Singh, S. V., Zimniak, P., Awasthi, S. and Awasthi, Y. C. (2001) Acceleratedmetabolism and exclusion of 4-hydroxynonenal through induction of RLIP76and hGST5.8 is an early adaptive response of cells to heat and oxidative stress.J. Biol. Chem. 276, 41213–4122342 Yang, Y., Cheng, J. Z., Singhal, S. S., Saini, M., Pandya, U., Awasthi, S. andAwasthi, Y. C. (2001) Role of glutathione S-transferases in protection againstlipid peroxidation. Overexpression of hGSTA2-2 in K562 cells protects againsthydrogen peroxide-induced apoptosis and inhibits JNK and caspase 3activation. J. Biol. Chem. 276, 19220–1923043 Sohal, R. S. and Orr, W. C. (1998) Oxidative stress may be a causal factor insenescence. Age 21, 81–8244 Parkes, T. L., Elia, A. J., Dickinson, D., Hilliker, A. J., Phillips, J. P. andBoulianne, G. L. (1998) Extension of Drosophila lifespan by overexpressionof human SOD1 in motorneurons. Nat. Genet. 19, 171–17445 Seong, K.-H., Ogashiwa, T., Matsuo, T., Fuyama, Y. and Aigaki, T. (2001) Applicationof the gene search system to screen for longevity genes in Drosophila.Biogerontology 2, 209–21746 Kang, H. L., Benzer, S. and Min, K. T. (2002) Life extension in Drosophila by feedinga drug. Proc. Natl. Acad. Sci. U.S.A. 99, 838–84347 Nanduri, B., Hayden, J. B., Awasthi, Y. C. and Zimniak, P. (1996) Amino acid residue104 in an alpha-class glutathione S-transferase is essential for the high selectivityand specificity of the enzyme for 4-hydroxynonenal. Arch. Biochem. Biophys. 335,305–31048 Armitage, P. and Berry, G. (1987) Statistical Methods in Medical Research. BlackwellScientific Publications, Oxford49 Felsenstein, J. (1993) PHYLIP (Phylogeny Inference Package), version 3.5c,Department of Genetics, University of Washington, Seattle, WAReceived 14 August 2002/8 November 2002; accepted 20 November 2002Published as BJ Immediate Publication 20 November 2002, DOI 10.1042/BJ20021287 2003 Biochemical Society